Mining Twitter Multi-word Product Opinions with Most Frequent Sequences of Aspect Terms

Given a corpus of microblog texts from a social media platform such as Twitter (e.g., “the new iPhone battery life is good, but camera quality is bad”), mining multi-word aspects (e.g., battery life, camera quality) and opinions (e.g., good, bad) of these products is challenging due to the vast amount of data being generated. Aspect-Based Opinion Mining (ABOM) is thus a combination of automatic aspect extraction and opinion mining that allows an enterprise to analyze the data on relevant features of products in detail, saving time and money. Existing Twitter ABOM systems such as Hate Crime Twitter Sentiment (HCTS) and Microblog Aspect Miner (MAM) generally go through the four-step approach of obtaining microblog posts, identifying frequent nouns (candidate aspects), pruning the candidate aspects, and getting opinion polarity. However, they differ in how well they prune their candidate features. This paper proposes a system called Microblog Aspect Sequence Miner (MASM) as an extension of Microblog Aspect Miner (MAM) by replacing the Apriori algorithm with a modified frequent sequential pattern mining algorithm based on CM-SPAM to also enable mining multi-word aspects more efficiently. The proposed system is able to determine the summary of most common aspects (Aspect Category) and their sentiments for a product. Experimental results with evaluation metrics of execution time, precision, recall, and F1-measure indicate that our approach has higher recall and precision than these existing systems on Sanders Twitter corpus dataset